no code implementations • 21 Jun 2019 • David Nass, Boris Belousov, Jan Peters
With the increasing pace of automation, modern robotic systems need to act in stochastic, non-stationary, partially observable environments.
Policy Gradient Methods